Unsolved
This post is more than 5 years old
7 Posts
0
5312
January 25th, 2012 07:00
symclone -cycled option
Hi,
we use clone + SRDF/S +snapshot on R2 on DMX4 in order to replicate our core ORACLE databases our scripts run the folowwing steps :
a) re-create clone with command symclone -cg xxxx recreate -precopy -i 300 -c 120
b) verify status with command symclone -cg xxxx verify -precopy -cycled -i 300 -c 120
c) put databases in backup mode
d) activate clone with command symclone -cg xxxx activate -consistent -i 300 -c 120
e) remove database from backup mode
f) terminate snapshot on R2
g) create and activate snpshot on R2
h) establish SRDF ( R1 are the clone devices)
i) check that CG is syncronized
j) split SRDF
now some questions
1) I would like to know is this methodogy is optimal
2) what the signification of the -cycled option ( I dont really undersand the explaination done on the man pages)
3) is this option mandatory in our configuration ?
thanks
dynamox
9 Legend
•
20.4K Posts
0
January 27th, 2012 22:00
2) see emc175391
3) it does not hurt, especially if you add new devices to an existing cg.
any reason you are using consistency group and not a regular device group ?
aatlan
7 Posts
0
January 30th, 2012 03:00
Paul,
thanks a lot for the explaination
It was very helpful
rawstorage
3 Apprentice
•
420 Posts
2
January 30th, 2012 03:00
Aharon,
The cycled switch on the verify command checks to see if all changed tracks have been copied to the target Device at least once since the recreate command was issued, when the cycled = True all changed tracks have been copied to the target, the precopy then cyles around again to see if any tracks have changed again since the last time they were copied, if they have changed again they are transferred to the target. This process continues in the background until the activate is issued. Verifying if the precopy has cycled can circumvent any potential performance issues seen on the R1 site when an activate is issued, so this is generally a good thing.
At the point of activate the bitmap of changed tracks is frozen and this is the point in time image that will be seen on your target. In your case you are doing a consistent activate which means host writes are held for a very brief period so that a write consistent image is captured across all devices in the group. This is a recommended best practice and gurantees the image.
I note you have opted to place the database in hot backup mode. This is no longer required by Oracle, as the consistent option guarantees a crash consistent image, however a lot of customers choose to still do this. In case of hot backup mode we generally recommend splitting the data and log devices into 2 distinct groups one for data one for logs, and activate the data group first while database is still in hot backup. Once data group is activated take out of hot backup mode and do a log switch then activate the clone of your log devices this gurantees the latest copy of the log file is on the clone so that database can verify when it was taken out of hot backup.
See the following techbook.
http://powerlink.emc.com/km/live1/en_US/Offering_Technical/White_Paper/h2603-oracle-db-emc-symmetrix-stor-sys-wp-ldv.pdf
The difference between a copy with the consistent option and hotbackup mode in the manner I have described restartable vs recoverable. This is also covered in detail in the techbook..
Hope this was helpful.
zellimbo
1 Message
0
October 10th, 2015 19:00
Useful blog post ! I Appreciate the points . Does someone know if my business can find a sample a form example to complete ?